The authors report about the surface modification of polystyrene (PSt) with photoreactive a-4-azidobenzoyl-w-methoxy poly(ethylene glycol)s (ABMPEG) of three different mol. wts. (MWs of .apprx.2, .apprx.5, and .apprx.10 kg/mol) and with two poly(ethylene glycol)/poly(propylene glycol) triblock copolymers (PEG-PPG-PEG) of about identical PEG/PPG ratio (80/20, wt./wt.) and MWPEG of .apprx.3 and .apprx.6 kg/mol, all via adsorption from aq. solns. For ABMPEGs, an addnl. UV irradn. was used for photografting to the PSt. Contact angle (CA) and at. force microscopy data revealed pronounced differences of the hydrophilicity/hydrophobicity and topog. of the surfaces as a function of PEG type and concn. used for the modification. In all cases, an incomplete coverage of the PSt was obsd. even after modification at the highest soln. concns. (10 g/L). However, clear differences were seen between PEG-PPG-PEGs and ABMPEGs; only for the latter was a nanoscale-ordered interphase structure with an influence of MWPEG on the PEG d. obsd.; after modification at the same soln. concns., the d. was significantly higher for lower MWPEG. The adsorption of three proteins, myoglobin (Mgb), bovine serum albumin (BSA), and fibrinogen to the various surfaces was analyzed by surface plasmon resonance. Pronounced differences between the two PEG types with respect to the redn. of protein adsorption were found. At high, but still incomplete, surface coverage and similar CA, the shielding of ABMPEG layers toward the adsorption of Mgb and BSA was much more efficient; e.g., the adsorbed Mgb mass relative to that of unmodified PSt was reduced to 10% for ABMPEG 2 kg/mol while for both PEG-PPG-PEGs the Mgb mass was still around 100%. In addn., for the ABMPEG layers an effect of MWPEG on adsorbed protein mass-decrease with decreasing MW-could be confirmed; and the highest Mgb/BSA selectivities were also obsd. A \"two-dimensional mol. sieving\", based on PEG mols. having a nanoscale order at the hydrophobic substrate polymer surface has been proposed, and the main prerequisites were the use of PEG conjugates which are suitable for an \"end-on\" grafting (e.g., ABMPEGs), the use of suitable (not too high) concns. for the surface modification via adsorption/self-assembly, optionally the photografting on the substrate (possible only for ABMPEG), and presumably, a washing step to remove the excess of unbound PEGs. The results of this study also strongly support the hypothesis that the biocompatibility of hydrophobic materials can be very much improved by PEG modifications at surface coverages that are incomplete but have an ordered layer structure controlled by the size and steric interactions of surface-bound PEGs.